J. Kim1, A. Watras1, H. Liu1, Z. Zeng1, J. A. Greenberg2, C. P. Heise2, Y. Hu1, H. Jiang1 1University Of Wisconsin-Madison,Electrical And Computer Engineering,Madison, WI, USA 2University Of Wisconsin School Of Medicine And Public Health,Department Of Surgery,Madison, WI, USA
Introduction: Laparoscopic surgery has numerous clinical benefits for patients including decreased postoperative pain, decreased wound morbidity, earlier recovery, and improved cosmesis. The quality and extent of intra-abdominal visualization is crucial during laparoscopic procedures. However, the current system that uses a single camera has many drawbacks in terms of the quality and extent of visualization as well as operative efficiency. We created a multi-camera visualization system with a larger field of view (FoV) compared to standard laparoscopic cameras for improved operative efficiency.
Methods: Our multi-camera visualization system utilizes four cameras with a specially designed trocar and a real-time video stitching program. We performed a bean drop task in a commercially available simulator box to test our large-FoV visualization system. Our video stitching program used the video data from four individual cameras and combined these four videos in real-time to provide large FoV. A projective transformation was calculated using matched feature points for mapping each image from different cameras into a single coordinate system. One of the four cameras was considered to provide the main view and then the required transformations were computed to map the images from four different cameras to the coordinate system of the main view.
Results: The four cameras are deployed and retrieved by the mechanical system of the trocar. The working port is not occupied by the cameras during operation and allows insertion of surgical instruments. Therefore, our system can reduce the number of ports or free up a surgical port, a potential advantage of our system. As a demonstration, we successfully performed a bean drop task in the trainer box by using our large-FoV visualization system. The figure below shows that a bean was picked up by a grasper and dropped into the hole of an inverted cup. The red dotted circles are traces of a moving bean picked by a grasper. The real-time stitched video can track the complete motion of the bean without any physical camera maneuver (Figure (a, b)), which is another clear advantage over current laparoscopic cameras that need to be operated separately by an assistant. In contrast, an image from a single camera cannot track the whole trajectory of the bean and cover the beans and the cup simultaneously (Figure (c, d)).
Conclusion: Our multi-camera visualization system provides a large FoV, frees up a surgical port, and eliminates the needs of physical maneuvering of the laparoscopic camera (thus the camera operating assistant). This system can provide larger extent of intra-abdominal visualization and may lead to higher efficiency of operation.
